62 IJCSNS International Journal of Computer Science and Network Security, VOL.14 No.

2, February 2014

The Methodology of Expert Systems

Kantureeva
Zakirova Alma Mannapova Torgyn Mussaif Marzhan Nigmetov Kanat
Mansiya

L.N.Gumilyov L.N. Gumilyov Zhangir-han West- S.Seifullin Kazakh L.N. Gumilyov

Eurasian National Eurasian National Kazakhstan agrarian- agrarian-technical Eurasian National
University University technical university university University

Summary activity. Development of expert systems is very different

The article considers design methodology of expert systems,
structure of the base components, studies stages of expert
system's creation, provides basic data on the ways of
representation of knowledge in expert systems.
Key words:
expert systems, knowledge database, software, intelligent
interface
1. Introduction
Nowadays there are no uniform methods, models and
tools for creating expert systems. More accessible
software tools are required, which would support a
standard design technology application of expert systems,
allow detailed analysis at the stage of systems design or
even bring a negative result. The purpose of the study is to
develop programs that get results that are not inferior in
quality and efficiency with the solutions obtained by an
expert. Scientific innovation is reflected in an attempt to
build a common methodology for the design of expert
systems, allowing creating cluster expert systems to solve
problems in a variety of subject areas. The practical
relevance of these results is to implement a method of
constructing intelligent expert system. The basic idea is to
create theoretical, technological and organizational
development techniques of expert systems.
from conventional software development.
The software tools that are based on the technology and
techniques of artificial intelligence greatly expand the
range of practically important problems that can be solved
by computers, and their solution brings significant
economic benefits. First, there is a problem that requires
combining different techniques and technologies into a
single coherent system, in which all the means and tools
must be used optimally. Second, the problems of the
single representation of knowledge in the integrated
environment remain unsolved; there are no methods for
organizing data exchange processes between expert
systems that are parts of it, no query processing
techniques to expert systems. There is no methodology of
designing an expert system as a whole. An overwhelming
number of expert systems, which were developed by the
classical architecture, are integral. Therefore, there is a
situation where you need to make changes in the
composition of the expert system, redesign expert systems,
change methods of handling requests, etc., which often
requires changes in the system and leads to a deadlock.
The experience of creating expert systems shows that the
use of the methodology adopted in the traditional
programming overly delays the process of creating expert
systems. Thirdly, there is no precise method of designing
expert systems.

2. The main problem of creating expert
systems technology
The expert system is a system of artificial intelligence,
which is built on the basis of deep expertise of a certain
subject area (obtained from experts and specialists in the
area). Expert systems are one of the few types of artificial
intelligence systems that are widely used and have found
practical application. Expert systems are a progressive
trend in the field of artificial intelligence. The increased
interest is determined by the possibility of their
application to problems of different fields of human
3. Technology of the creating expert systems
Expert systems are one of the fields of artificial
intelligence (Artificial Intelligence - AI). Research in this
area is focused on the development and implementation of
computer programs that can emulate (imitate, reproduce)
those human activities that require thinking, some skills
and gathered experience. These include the problem of
decision-making [1], image recognition and understanding
of human language.

Manuscript received February 5, 2014

Manuscript revised February 20, 2014
IJCSNS International Journal of Computer Science and Network Security, VOL.14 No.2, February 2014
Decision-
maker
User Knowledge Knowledge
Interface Base Base Editor
Explanation
System
User Engineer/Expert
Figure 1. The typical structure of an expert system
Development of expert systems is possible for a particular
application if the following requirements are met:
1) there are experts in this field, who solve the
problem much better than the novice
professionals ;
2) the experts agree on the proposed assessment
solutions, otherwise it is impossible to evaluate
the quality of developed ES;
3) the experts are able to express in natural
language and explain the methods used by them,
otherwise it is difficult to count on the fact that
the knowledge of experts will be "removed" and
put in the ES;
4) the solution of the problem requires only the
reasoning, not action;
5) the problem should not be too difficult (i.e., the
decision making should take a few hours or
days for expert);
6) the problem should be "clear" enough and
structured, i.e. there should be marked basic
concepts, relationships, and well-known (at
least for expert) methods of obtaining solutions
of the problem;
The use of expert systems in the application is possible
and justified by one of the following factors:
the solution will bring a significant effect, such as
economic;
the use of a human-expert is not available either because
of lack of experts or the need to carry out the examination
at the same time in different places;
63
the use of ES is appropriate in those cases where during
the transfer of information to expert unacceptable loss of
time or information occurs.
Herein solvable problem must have a set of the
following characteristics:
1) The problem can be solved by a natural manipulation
of symbols (i.e., using symbolic arguments), and not by
manipulation of numbers, as is customary in mathematical
techniques and traditional programming;
2) The problem must have a heuristic nature, and not
algorithmic, i.e. its decision shall require the use of
heuristics rules;
3) The problem must repay expenditures of developing ES.
However, it should not be too difficult (decision making
takes hours for expert, but not weeks), so ES can solve it;
4) The problem should be practically significant.
4. The scheme of expert system design
In general, the expert system consists of the following
typical subsystems: knowledge acquisition, knowledge
management classification as belonging to a local expert
systems, knowledge management subsystem output,
output management expertise in a cluster expert system
using a mathematical model for constructing a tree
inference; subsystems forming opinions and
recommendations, a set of local expert systems derived
from the decomposition of knowledge chosen subject area.
A distinguishing feature of expert systems is their ability
to accumulate knowledge and experience of the most
qualified specialists (experts) in a narrow subject area.
Then with the help of this knowledge ordinary skilled
users with the help of expert systems can solve their
current problems as successfully as it would make the
experts themselves. This effect is achieved due to the fact
that the expert system in their work plays roughly the
same line of argument, which typically uses a human
expert in the analysis of a problem. Thus, expert systems
allow you to copy and distribute knowledge, making the
unique experience of high quality professionals available
to a wide circle of ordinary skills. Level of expert systems
users may vary over a wide range. Functions that are
applied in the users expert systems also depend on their
activities.
Expert systems are built on the three "pillars"[3]:
Knowledge gained from the experts;
Mapping method in the knowledge database;
Methods of using the knowledge databases for logic
conclusion.
64 IJCSNS International Journal of Computer Science and Network Security, VOL.14 No.2, February 2014
Figure 2. The scheme of the expert system
A list of typical problems, which can be solved by expert
systems, includes [2] following:
Extracting information from the raw data (such as
signals from the sonar);
Fault diagnostics (as in technical systems as well as in
humans);
Structural analysis of complex objects (for example,
chemical compounds);
Selecting the configuration of complex multi-component
systems (e.g. distributed computing systems)
Figure 3. The architecture of an expert system
Knowledge database is a special kind of database
designed to manage the metadata, which is, to collect,
store, search, and deliver knowledge. Knowledge
databases are often applied in the context of expert
systems, where they are used to represent the skills and
experience of experts engaged in practical activities in the
related field (such as medicine or mathematics).
Inference mechanism - this software tool receives the
request converted into internal representation from
intellectual interface, forms a specific algorithm of
problem solution from knowledge database, and performs
an algorithm. The intelligent interface then obtains the
received result for a response to a request.
Mechanism of knowledge acquisition is below. The
knowledge database reflects experts knowledge
(specialists) in this area of concern about the actions in
various situations or processes of specific task solutions.
Identification of such knowledge and subsequent
submission to the knowledge database is performed by
experts called knowledge engineers.
Mechanism of knowledge explanation - in process of
solving or according to the results of the problem, user
can request an explanation or justification of the solving
process. For this purpose ES should provide an
appropriate explanation mechanism.
Intelligent interface - is the user interface that is
additionally equipped with software that can perform
basic functions of analysis, synthesis, comparison,
compilation, storage, and training of all components
involved in the process of interaction with the user.
5. Technology of expert systems development
Technology of expert systems development includes six
following steps (Fig. 2): identification, conceptualization,
formalization, implementation, testing, and trial operation.
On the stage of identification the development process of
a prototype system is planned [4], the sources of
knowledge (books, experts, and methodologies), goals
(distribution of experience, automation of routine
operations), classes of problems solved and other things
are determined. The result of identification is a response
to a question of what is needed to be done and what
resources are required to be used.
During the stage of conceptualization profound analysis
of the problem domain is conducted, the definitions and
their relationships are identified, and methods of problem
solving are defined. This stage is completed by the
creation of the domain model (software) that includes
basic concepts and relationships [5]. The following
problem characteristics are defined during the stage of
conceptualization:
IJCSNS International Journal of Computer Science and Network Security, VOL.14 No.2, February 2014 65

experts actions towards the real problem solutions. The

purpose of this stage is the creation of one ES prototype.
Subsequently after the testing results and trial operation a
final product that is suitable for industrial use is created at
this stage [8]. Prototype development consists of its
components programming or their selection among
famous software tools and filling the knowledge database
[9].
Thus, nowadays the process of situation assessment and
decision-making is one the most time-consuming and,
therefore, the expert systems development methodology
considers following necessities to be the most important
concern:
1) Correct formulation of the problem;
2) Knowledge systemization for its transfer to the
computer system;
3) The development of management tools for knowledge
database, inductive inference and simplified dialogue
interaction.

Figure 4. Methods of expert systems development

Types of data available;

Source and output data, sub problems of general
problem;
Applied strategies and hypotheses;
Types of relationships between software objects, types
of used relations (hierarchy, cause - effect, part - whole,
etc.);
Processes used in the solutions;
Knowledge content used in problem solution;
Types of constraints imposed on the processes used in
the solutions;
Knowledge content used in the decision explanations [5].

During the stage of formalization IP is selected and ways

of presentation of all kinds of knowledge are determined,
basic concepts are formalized, ways of knowledge
interpretation are defined, system operation is simulated,
and adequacy of the objectives concerning the system of
fixed definitions, solving methods, ways of presentation
and knowledge manipulation is evaluated [6]. The output
of formalization stage is a description of how the problem
can be represented in the selected or developed formalism
(frames, scripts, semantic networks, etc.) and an
understanding of manipulation methods of this knowledge
(the logical conclusion) [7].
During the performance stage which is filling database Figure 5. Stages of expert systems development
with expert. The process of acquiring knowledge is
performed by a knowledge engineer on the basis of
66 IJCSNS International Journal of Computer Science and Network Security, VOL.14 No.2, February 2014

References: Kantureyeva Mansiya the senior teacher

[1] Luger D.F. Artificial Intelligence: Strategies and methods of information systems department of L.N.
to solve difficult problems / DF Luger. - 4th Ed. - Moscow: Gumilyov Eurasian National University.
Publishing House "Williams", 2005. - 863 p. - ISBN 5-
8459-0437-4: 489.28. - 489.28.
[2] Russell, S. Artificial Intelligence: A modern Approach / S.
Russell, P. Norvig. - 2nd Ed. - Moscow: Publishing House
Williams, 2007. - 1407 p. - (The Book of intelligent
agents). - ISBN 978-5-8459-0887-2: 666 p.
[3] Devyatkov, V.V. Artificial Intelligence Systems: Study Zakirova Alma the senior teacher of
manual for higher education institutions / V.V. Devyatkov; computer science department of L.N.
Chief Editor I.B.Fedorov. - Moscow: Publishing House of Gumilyov Eurasian National University.
the Bauman Moscow State Technical University, 2001. -
351 p.
[4] Chastikov A.P. Gavrilova T.A. Belov D. L. Development
of expert systems. CLIPS environment. SPb. : BHV-
Petersburg, 2003. 608 p.
[5] Informatics: The textbook / Under the editorship of the prof.
of N. V. Makarova M: Finance and statistics, 1997. Mannapova Torgyn the senior teacher of computer science of
[6] Artificial intelligence: Application in the integrated Zhangir-han West-Kazakhstan agrarian-technical University.
production systems / Under the editorship of E.Kyyusiak;
The Lane with English M: Mechanical engineering, 1991. Mussaif Marzhan the senior teacher of computer science of
[7] Artificial intelligence: in 3 books of Book 1. Systems of S.Seifullin Kazakh agrarian-technical University.
communication and expert systems. The directory / Under
the editorship of E.V.Popova M: Radio and Nigmetov Kanat the senior teacher of computer science of L.N.
communication, 1990. Gumilyov Eurasian National University.
[8] Page Osuga. Processing of knowledge: The lane with .
M: World, 1989.
[9] Priests E.V.Ekspertnye systems: The solution of
unformalized tasks in dialogue with the COMPUTER. M:
Science. Hl. edition physical. 1987 .